tcp westwood (with faster recovery) claudio casetti ([email protected]) mario gerla...

20
TCP Westwood (with Faster Recovery) Claudio Casetti ([email protected]) Mario Gerla ([email protected]) Scott Seongwook Lee ([email protected]) Saverio Mascolo ([email protected]) Medy Sanadidi ([email protected]) Computer Science Department University of California, Los Angeles, USA

Post on 21-Dec-2015

220 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: TCP Westwood (with Faster Recovery) Claudio Casetti (casetti@polito.it) Mario Gerla (gerla@cs.ucla.edu) Scott Seongwook Lee (sslee@cs.ucla.edu) Saverio

TCP Westwood (with Faster Recovery)

Claudio Casetti ([email protected])

Mario Gerla ([email protected])

Scott Seongwook Lee ([email protected])

Saverio Mascolo ([email protected])

Medy Sanadidi ([email protected])

Computer Science Department

University of California, Los Angeles, USA

Page 2: TCP Westwood (with Faster Recovery) Claudio Casetti (casetti@polito.it) Mario Gerla (gerla@cs.ucla.edu) Scott Seongwook Lee (sslee@cs.ucla.edu) Saverio

TCP Congestion Control • Based on a sliding window algorithm• Two stages:

– Slow Start, initial probing for available bandwidth (“exponential” window increase until a threshold is reached)

– Congestion Avoidance,”linear” window increase by one segment per RTT

• Upon loss detection (coarse timeout expiration or duplicate ACK) the window is reduced to 1 segment (TCP Tahoe)

Page 3: TCP Westwood (with Faster Recovery) Claudio Casetti (casetti@polito.it) Mario Gerla (gerla@cs.ucla.edu) Scott Seongwook Lee (sslee@cs.ucla.edu) Saverio

Congestion Window of a TCP Connection Over Time

Page 4: TCP Westwood (with Faster Recovery) Claudio Casetti (casetti@polito.it) Mario Gerla (gerla@cs.ucla.edu) Scott Seongwook Lee (sslee@cs.ucla.edu) Saverio

Shortcomings of current TCP congestion control

• After a sporadic loss, the connection needs several RTTs to be restored to full capacity

• It is not possible to distinguish between packet loss caused by congestion (for which a window reduction is in order) and a packet loss caused by wireless interference

• The window size selected after a loss may NOT reflect the actual bandwidth available to the connection at the bottleneck

Page 5: TCP Westwood (with Faster Recovery) Claudio Casetti (casetti@polito.it) Mario Gerla (gerla@cs.ucla.edu) Scott Seongwook Lee (sslee@cs.ucla.edu) Saverio

New Proposal:TCP with “Faster Recovery”

• Estimation of available bandwidth (BWE):– performed by the source– computed from the arrival rate of ACKs,

smoothed through exponential averaging

• Use BWE to set the congestion window and the Slow Start threshold

Page 6: TCP Westwood (with Faster Recovery) Claudio Casetti (casetti@polito.it) Mario Gerla (gerla@cs.ucla.edu) Scott Seongwook Lee (sslee@cs.ucla.edu) Saverio

TCP FR: Algorithm Outline

• When three duplicate ACKs are detected:– set ssthresh=BWE*rtt (instead of ssthresh=cwin/2 as in Reno)

– if (cwin > ssthresh) set cwin=ssthresh

• When a TIMEOUT expires:– set ssthresh=BWE*rtt (instead of ssthresh=cwnd/2 as in Reno) and cwin=1

Page 7: TCP Westwood (with Faster Recovery) Claudio Casetti (casetti@polito.it) Mario Gerla (gerla@cs.ucla.edu) Scott Seongwook Lee (sslee@cs.ucla.edu) Saverio

Experimental Results

• Compare behavior of TCP Faster Recovery with Reno and Sack

• Compare goodputs of TCP with Faster Recovery, TCP Reno and TCP Sack– with bursty traffic (e.g., UDP traffic)– over lossy links

Page 8: TCP Westwood (with Faster Recovery) Claudio Casetti (casetti@polito.it) Mario Gerla (gerla@cs.ucla.edu) Scott Seongwook Lee (sslee@cs.ucla.edu) Saverio

FR/Reno Comparison

0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

0.16

0 100 200 300 400 500 600 700 800

no

rma

lize

d t

hro

ug

hp

ut

Time (sec)

1 TCP + 1 On/Off UDP (ON=OFF=100s) 5 MB buffer - 1.2s RTT - 150 Mb/s Cap.

FR

Reno

Page 9: TCP Westwood (with Faster Recovery) Claudio Casetti (casetti@polito.it) Mario Gerla (gerla@cs.ucla.edu) Scott Seongwook Lee (sslee@cs.ucla.edu) Saverio

Goodput in presence of UDPDifferent Bottleneck Sizes

0

1

2

3

4

5

6

0 20 40 60 80 100 120 140 160

Goo

dput

[M

b/s]

Bottleneck bandwidth [Mb/s]

FR

RenoSack

Page 10: TCP Westwood (with Faster Recovery) Claudio Casetti (casetti@polito.it) Mario Gerla (gerla@cs.ucla.edu) Scott Seongwook Lee (sslee@cs.ucla.edu) Saverio

Wireless and Satellite Networks

0

200000

400000

600000

800000

1e+06

1.2e+06

1.4e+06

1e-10 1e-09 1e-08 1e-07 1e-06 1e-05 0.0001 0.001

good

put

(bits

/s)

bit error rate (logscale)

TahoeRenoFR

link capacity = 1.5 Mb/s - single “one-hop” connection

Page 11: TCP Westwood (with Faster Recovery) Claudio Casetti (casetti@polito.it) Mario Gerla (gerla@cs.ucla.edu) Scott Seongwook Lee (sslee@cs.ucla.edu) Saverio

Experiment Environment

Router

Source

Linkem ulator

D estination

Source

100M

100M

• New version of TCP FR called “TCP Westwood”•TCP Westwood is implemented in Linux kernel 2.2.10.• Link emulator can emulate:

• link delay• loss event

• Sources share bottleneck through router to destination.

Page 12: TCP Westwood (with Faster Recovery) Claudio Casetti (casetti@polito.it) Mario Gerla (gerla@cs.ucla.edu) Scott Seongwook Lee (sslee@cs.ucla.edu) Saverio

Goodput Comparison with Reno (Sack)

goodput vs. pipesize

0

1

2

3

0.02 0.27 0.52 1.02 2.52 5.02

pipesize(Mb)

good

put(M

b/Se

c)

Westwood

Reno withSack

• Bottleneck capacity 5Mb• Packet loss rate 0.01• Larger pipe size

corresponds to longer delay

goodput vs. loss rate

0

0.5

1

0 0 0 0 0

Packet Loss Rate

good

put

(Mb/

Sec) Westwood

Reno withSack

• Link delay 300ms• Bottleneck bandwidth 5Mb• Concurrent on-off UDP

traffic

Page 13: TCP Westwood (with Faster Recovery) Claudio Casetti (casetti@polito.it) Mario Gerla (gerla@cs.ucla.edu) Scott Seongwook Lee (sslee@cs.ucla.edu) Saverio

Friendliness with Reno

• Goodput comparison when TCP-W and Reno share the same bottleneck– over perfect link– 5 Reno start first– 5 west start after 5 seconds– 100 ms link delay

• Goodput comparison when TCP-W and Reno share the same bottleneck– over lossy link(1%)– 3 Reno start first then 2 Westwood– 100 ms link delay

• TCP-W improves the performance over lossy link but does not catch the link.

Page 14: TCP Westwood (with Faster Recovery) Claudio Casetti (casetti@polito.it) Mario Gerla (gerla@cs.ucla.edu) Scott Seongwook Lee (sslee@cs.ucla.edu) Saverio

Current Status & Open Issues

• Extended testing of TCP WEswoh

• Friendliness/greediness towards other TCP schemes

• Refinements of bandwidth estimation process

• Behavior with short-lived flows, and with large number of flows

Page 15: TCP Westwood (with Faster Recovery) Claudio Casetti (casetti@polito.it) Mario Gerla (gerla@cs.ucla.edu) Scott Seongwook Lee (sslee@cs.ucla.edu) Saverio

Extra slides follow

Page 16: TCP Westwood (with Faster Recovery) Claudio Casetti (casetti@polito.it) Mario Gerla (gerla@cs.ucla.edu) Scott Seongwook Lee (sslee@cs.ucla.edu) Saverio

Losses Caused by UDPDifferent RTT

0.8

1

1.2

1.4

1.6

1.8

2

2.2

2.4

2.6

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

Goo

dput

[M

b/s]

one-way RTT (s)

FR

RenoSack

Page 17: TCP Westwood (with Faster Recovery) Claudio Casetti (casetti@polito.it) Mario Gerla (gerla@cs.ucla.edu) Scott Seongwook Lee (sslee@cs.ucla.edu) Saverio

Losses Caused by UDPDiffererent Number of Connections

0

1

2

3

4

5

6

7

8

9

10

11

0 5 10 15 20 25 30

Goo

dput

[M

b/s]

no. of connections

FR1

RenoSack

Page 18: TCP Westwood (with Faster Recovery) Claudio Casetti (casetti@polito.it) Mario Gerla (gerla@cs.ucla.edu) Scott Seongwook Lee (sslee@cs.ucla.edu) Saverio

TCP over Lossy linksDifferent Bottleneck Size

0.1

1

10

0 20 40 60 80 100 120 140 160

Goo

dpu

t [M

b/s]

Bottleneck bandwidth [Mb/s]

FR

RenoSack

Page 19: TCP Westwood (with Faster Recovery) Claudio Casetti (casetti@polito.it) Mario Gerla (gerla@cs.ucla.edu) Scott Seongwook Lee (sslee@cs.ucla.edu) Saverio

Bursty trafficdiffererent number of connections

0

2

4

6

8

10

12

14

0 5 10 15 20 25 30

Goo

dpu

t [M

b/s]

no. of connections

FRRenoSack

Page 20: TCP Westwood (with Faster Recovery) Claudio Casetti (casetti@polito.it) Mario Gerla (gerla@cs.ucla.edu) Scott Seongwook Lee (sslee@cs.ucla.edu) Saverio

Fairness of TCP Westwood

• Cwnds of two TCP Westwood connections– over lossy link– concurrent UDP traffic– timeshifted– link delay 100ms

• Concurrent TCP-W connections goodput– 5 connections (other2 are

similar)– link delay 100ms.